The emergence of cloud-based database systems has intensified the challenge of achieving optimal performance under large-scale workloads. Conventional tuning methods, which rely on human intervention and static heuristics, often struggle with the complexity and diversity of modern cloud environments. This survey offers an in-depth examination of the reinforcement learning (RL) techniques of automated optimisation of cloud database performance. By continuously observing workload behavior and system feedback, RL enables the database engines to become self-adaptive and reduces the need for continuous manual monitoring.
For more than 30 years, the IBM AS/400 (iSeries/IBM i) platform has supported mission-critical operations in retail and banking with unmatched reliability, seamless database management, and high performance. These legacy environments, however, are under tremendous strategic pressure amid the rapid pace of digital transformation, cloud-native architectures, and the demands of the API economy. In this paper, an in-depth, evidence-based approach is presented to modernize enterprise systems built on the AS/400 platform, supported by case studies of 12 retail chains and 8 banking institutions from North America and Europe.
As large language models (LLMs) continue to mature quickly, the focus of enterprise AI has shifted from simply leveraging the model to the integration, orchestration, and governance of enterprise AI. While the shift from chatbots to agentic ones that reason, plan, and make decisions on their own across a plethora of different data sources and tools, there are no universal frameworks that connect models to enterprise resources. In response to this challenge, this review discusses the Model Context Protocol (MCP) and the entire class of agentic AI middleware, in the context of enterprise integration and the development of middleware. We combine existing work from the areas of reasoning, tool use, retrieval augmented grounding, multi-agent coordination and AI security to extract the key concepts and research gaps of this field.
Energy-conscious bare-metal provisioning has become an important research question since virtualized data centers are now full of mixed workloads, whose performance, isolation, and energy usage vary radically across virtual machines, containers, and direct hardware allocation.
As the number of microservices increases, developer platforms are increasingly expected to support scalability by incorporating command line interfaces with standard delivery, observability and governance capabilities into the backend workflows of software development teams.
Digital content is an important sales enabler and AI-augmented CMSs play a crucial role in making content easily accessible for sales teams, relevant to sales teams’ needs and buyer contexts, personalized, compliant, measurable and revenue-generating.
As new and multiple cloud platforms emerge, each providing different features and capabilities, the data journey is becoming more complex and more important, resulting in an autonomous multi-cloud data pipeline orchestration problem such as container platforms, serverless services, streaming engines, data lakes, and disparate cloud providers host many analytical workloads.
AI coding agents using large language model (LLM) technology are revolutionizing the software development landscape, especially in the areas of planning, coding, testing, deployment, and maintenance.
The success of the latest developments in multimodal artificial intelligence and large language models has made Design-to-Code automation even more efficient and straightforward for converting a written design document into working front-end application source code. With the current explosion of multimodal AI and the advent of generative large language models, the above-mentioned Design-to-Code automation initiative takes on new momentum by turning text design docs into usable application front-end code.
The success of the latest developments in multimodal artificial intelligence and large language models has made Design-to-Code automation even more efficient and straightforward for converting a written design document into working front-end application source code. With the current explosion of multimodal AI and the advent of generative large language models, the above-mentioned Design-to-Code automation initiative takes on new momentum by turning text design docs into usable application front-end code.
The aim of this review is to suggest the Agentic GTM Maturity Model as a simple matrix to evaluate how mature an AI-powered go-to-market revenue system and related organisations are in their transition from uncoordinated analytics to managed semi-autonomous revenue systems. The academic perspectives of the peer-reviewed journal articles reviewed herein build upon one another when it comes to understanding the agentic maturity of GTM and provide insights to AI-driven sales, B2B marketing automation, CRM intelligence, customer engagement, customer-retention analytics, and digital transformation.
Edge clouds are gaining momentum as a way to enable resource-constrained, latency-sensitive applications in smart cities, industrial automation, health care, and autonomous vehicles. With workloads being shared among heterogeneous devices and networks, it presents a complex performance monitoring/analysis problem. Conventional centralized monitoring systems may not be able to provide real-time insights at the level and pace that is needed in edge environments.
Quantum measurement involves a transition from a quantum description containing superposed alternatives to physical configurations that can function as stable classical records. Although measurement interaction, decoherence, environmental encoding, and Quantum Darwinism explain important aspects of this quantum-to-classical transition, they do not by themselves provide a unified criterion for determining when measurement information becomes a sufficiently stable and accessible physical record. This paper introduces Classical Event Anchoring (CEA), a record-centred conceptual framework that organizes the state-to-record transition through system–apparatus correlation, decoherence, record stability, communicability, and practical irreversibility.
Autonomous Cloud Database Performance Optimization Using Reinforcement Learning in AWS and Snowflake
Raghu Gopa
Download
The emergence of cloud-based database systems has intensified the challenge of achieving optimal performance under large-scale workloads. Conventional tuning methods, which rely on human intervention and static heuristics, often struggle with the complexity and diversity of modern cloud environments. This survey offers an in-depth examination of the reinforcement learning (RL) techniques of automated optimisation of cloud database performance. By continuously observing workload behavior and system feedback, RL enables the database engines to become self-adaptive and reduces the need for continuous manual monitoring.
Modernizing AS/400-Based Enterprise Systems: A Practical Framework for Retail and Banking Transformation
Shiva Kumar Devasani
Download
For more than 30 years, the IBM AS/400 (iSeries/IBM i) platform has supported mission-critical operations in retail and banking with unmatched reliability, seamless database management, and high performance. These legacy environments, however, are under tremendous strategic pressure amid the rapid pace of digital transformation, cloud-native architectures, and the demands of the API economy. In this paper, an in-depth, evidence-based approach is presented to modernize enterprise systems built on the AS/400 platform, supported by case studies of 12 retail chains and 8 banking institutions from North America and Europe.
Model Context Protocols and Agentic AI Middleware: An Enterprise Architecture Framework for Scalable AI-Driven
Digital Transformation
Amil Bhadreshkumar Shah
Download
As large language models (LLMs) continue to mature quickly, the focus of enterprise AI has shifted from simply leveraging the model to the integration, orchestration, and governance of enterprise AI. While the shift from chatbots to agentic ones that reason, plan, and make decisions on their own across a plethora of different data sources and tools, there are no universal frameworks that connect models to enterprise resources. In response to this challenge, this review discusses the Model Context Protocol (MCP) and the entire class of agentic AI middleware, in the context of enterprise integration and the development of middleware. We combine existing work from the areas of reasoning, tool use, retrieval augmented grounding, multi-agent coordination and AI security to extract the key concepts and research gaps of this field.
Energy-Aware Bare-Metal Provisioning in Virtualized Data Centers
Ratan Raj Anandeshi
Download
Energy-conscious bare-metal provisioning has become an important research question since virtualized data centers are now full of mixed workloads, whose performance, isolation, and energy usage vary radically across virtual machines, containers, and direct hardware allocation.
Design Patterns for Scalable Developer Platforms: A Microservices and CLI-Centric Approach
Shubham Srivastava
Download
As the number of microservices increases, developer platforms are increasingly expected to support scalability by incorporating command line interfaces with standard delivery, observability and governance capabilities into the backend workflows of software development teams.
Designing AI-Augmented Content Management Systems for Sales Enablement: A Framework for Measuring Engagement
to-Revenue Impact
Nisha Gopinath Menon
Download
Digital content is an important sales enabler and AI-augmented CMSs play a crucial role in making content easily accessible for sales teams, relevant to sales teams’ needs and buyer contexts, personalized, compliant, measurable and revenue-generating.
Autonomous Multi-Cloud Data Pipeline Orchestration Using AI-Driven Observability and Self-Healing ETL Frameworks
Sreenivasa Reddy Vemareddy
Download
As new and multiple cloud platforms emerge, each providing different features and capabilities, the data journey is becoming more complex and more important, resulting in an autonomous multi-cloud data pipeline orchestration problem such as container platforms, serverless services, streaming engines, data lakes, and disparate cloud providers host many analytical workloads.
Enhancing Cost Efficiency of AI Coding Agents at Enterprise Scale: Challenges, Strategies, and Frameworks
Sainadh Ainala, Vinay Chowdary Duvvada
Download
AI coding agents using large language model (LLM) technology are revolutionizing the software development landscape, especially in the areas of planning, coding, testing, deployment, and maintenance.
AI-Assisted Design-to-Code Automation for Enterprise Frontend Engineering
Mohammed Siddiq Hussain Nisar Khan
Download
The success of the latest developments in multimodal artificial intelligence and large language models has made Design-to-Code automation even more efficient and straightforward for converting a written design document into working front-end application source code. With the current explosion of multimodal AI and the advent of generative large language models, the above-mentioned Design-to-Code automation initiative takes on new momentum by turning text design docs into usable application front-end code.
Intelligent Failure Prediction in Enterprise Systems Using Behavioural Pattern Modelling
Pradip Baral
Download
The success of the latest developments in multimodal artificial intelligence and large language models has made Design-to-Code automation even more efficient and straightforward for converting a written design document into working front-end application source code. With the current explosion of multimodal AI and the advent of generative large language models, the above-mentioned Design-to-Code automation initiative takes on new momentum by turning text design docs into usable application front-end code.
The Agentic GTM Maturity Model: A Framework for AI-Driven Revenue Systems
Akhila Gudla
Download
The aim of this review is to suggest the Agentic GTM Maturity Model as a simple matrix to evaluate how mature an AI-powered go-to-market revenue system and related organisations are in their transition from uncoordinated analytics to managed semi-autonomous revenue systems. The academic perspectives of the peer-reviewed journal articles reviewed herein build upon one another when it comes to understanding the agentic maturity of GTM and provide insights to AI-driven sales, B2B marketing automation, CRM intelligence, customer engagement, customer-retention analytics, and digital transformation.
Performance Analytics with Grafana and Prometheus for Edge Cloud Architectures
Veeresh Nunavath
Download
Edge clouds are gaining momentum as a way to enable resource-constrained, latency-sensitive applications in smart cities, industrial automation, health care, and autonomous vehicles. With workloads being shared among heterogeneous devices and networks, it presents a complex performance monitoring/analysis problem. Conventional centralized monitoring systems may not be able to provide real-time insights at the level and pace that is needed in edge environments.
Classical Event Anchoring in Quantum Measurement: A Scale-Relative Record Criterion for the State-to-Record Transition
Niraj Pathak, Dinesh Kumar Garg, DhirajKumar Pathak
Download
Quantum measurement involves a transition from a quantum description containing superposed alternatives to physical configurations that can function as stable classical records. Although measurement interaction, decoherence, environmental encoding, and Quantum Darwinism explain important aspects of this quantum-to-classical transition, they do not by themselves provide a unified criterion for determining when measurement information becomes a sufficiently stable and accessible physical record. This paper introduces Classical Event Anchoring (CEA), a record-centred conceptual framework that organizes the state-to-record transition through system–apparatus correlation, decoherence, record stability, communicability, and practical irreversibility.